17th-century style science in the 21st century? The case of the Alzheimer's disease vaccine
European Pharmaceutical Review
In the year 1633, the Roman Catholic Inquisition decided after investigating the facts that the Heliocentric Theory was absurd and declared its champion the Italian polymath
In scientific areas dealing with living organisms, because of their complexity there should be a constant review of studies' assumptions and results. Consequently, areas like drug development have benefited by learning from prior errors, identifying their reasons and ways to resolve them. This requirement is more critical in new areas, lacking significant depositories of past information to tap. But, a rational approach to drug development does not mean focusing solely on one strategy while ignoring other alternatives. A process that benefits from the knowledge of different areas, such as drug development, is largely the result of a multidisciplinary scientific approach, rather than a monolithic one.
Evidently, the lessons from the advances in medicinal chemistry over the last 200 years have resulted in a revolution in drug development, with antibiotics being a good example of the success of the trial and error approach. Areas like cancer and antiviral drugs have also shown similar successful growth using the same approach. However, that long-term learning curve has not taken place in Alzheimer's disease (AD), where due to a sudden increase in the ageing population, a result of the success of modern medicine in extending our lifespan, humanity faces now an epidemic of a disease that was rare not long ago.
This critical situation is evident in the attempts to develop products based on immunological approaches, like vaccination and antibody therapy, to treat, prevent or delay the onset of AD. This strategy is supported by various reports describing the protective effects of natural anti-AD antibodies, which start to decline with ageing and as the disease begins to manifest.2-4
Evidence for aducanumab
But the most compelling evidence of this protective immunity is the monoclonal antibody aducanumab, which is a replica of an antibody against aberrant cytotoxic forms of amyloid ß, found in elderly but cognitively normal people.5 Indeed, it stops and improves the mental decline linked to Alzheimer’s, all signs that anticipate the success of immune therapy. Yet all of the studies using immunological approaches to prevent and/or treat Alzheimer’s disease, with the possible exception of aducanumab, have failed. This has been taken as proof that the immunological approach is unrealistic. But the failures in AD drug development are not limited to the vaccine approach, and extend to all of the other methods being explored.6
An outcome of the unsubstantiated conclusion that immunologically based methods targeting amyloid-ß are unfeasible is the tacit suppression of research concerning vaccines targeting this protein as a strategy against Alzheimer’s disease.7 A decision that, like that of the 17th-Century Roman Inquisition, assumes that the conclusions derived from dubious studies are the correct ones. Indeed, a review of the assumptions used to design the AD vaccine studies shows errors made from the discovery up to the clinical phase, including some decisions made by regulatory agencies.
Constant discrepancies should have been a warning about problems with the transgenic mouse models and the need to obtain confirmation using animal models
Recurrent discrepancies between studies that are supposed to agree implies that those studies as planned are conflicting. Such is the situation with AD immune therapy, where studies in transgenic mouse AD models consistently gave promising results, which warranted these products to enter the clinical phase where they failed. Transgenic animal models, while helpful to explain the pathogenesis of AD, offer only a partial view of this disease where several other important cofactors are missing.8 This explains why treatments that apparently improved some of the pathology in transgenic mice failed in humans where the disease is more intricate.
Yet those constant discrepancies should have been a warning about problems with the transgenic mouse models and the need to obtain confirmation using animal models, such as dogs and non-human primates where AD occurs with ageing similar to humans.9 Evidently those clinical failures were due to incomplete scientific planning rather than a scientifically flawed immunological approach.
However, the problems in AD vaccine development extend also to the vaccines themselves. A review shows that while these vaccines were intended to remove plaque, the crucial change in transgenic mouse models, they ignored the cytotoxic soluble forms of amyloid-ß responsible for the neurotoxicity observed in AD.7,10-11
Actually, the first vaccine that failed and harmed some AD patients, AN-1792, also included the potent inflammatory agent QS-21, despite the fact that inflammation worsens this disease. Paradoxically, the AN-1792 vaccine was tested in Phase II using a formulation that was different from the Phase I vaccine and that induced a potent inflammatory immune response causing brain damage. In fact, addition of the non-ionic detergent polysorbate to the AN-1792 vaccine during the Phase II caused a massive increase in the adjuvanticity of QS-21; an effect disclosed in patent application WO1999010008A1 published in 1999.12
Changes in adjuvanticity
While information from patents is not as thorough as that from scientific publications, the reported significant changes in adjuvanticity caused by the addition of polysorbate should have been a warning of potential problems. Thus, the complications with the AN-1792 vaccine were the result of amplifying the QS-21 undesirable inflammatory Th1 immunity by the addition of polysorbate. Actually, this was an unusual situation, because in clinical studies the formulation used in Phase II is always the one that was proven to be safe in Phase I. Nevertheless, in the AN-1792 case, those rules were evidently overlooked with unfortunate results; a bad outcome that while assumed to be the result of formulation problems also fostered questions about the vaccine concept.
A result of that crisis was the development of safer vaccines, but still using transgenic mouse models. Furthermore, these presumably safer vaccines frequently had strongly inflammatory adjuvants like QS-21 or similar ones13,14 - vaccines that were successful in mouse models but failed in humans. The fact that the elicited immune response after vaccination depends solely on the adjuvant, regardless of the antigen's nature, was apparently ignored in these vaccines. Hence, the negative results were most likely due to the wrong vaccine formulations, which, rather than eliciting a protective anti-AD immunity, induced a detrimental one.7
This confusion affecting Alzheimer’s disease vaccine development is shown by these vaccines' composition and the reported results. For instance, a recent article concludes that the long-term administration to humans of an Alzheimer’s disease vaccine having QS-21 seems to be safe,15 despite the fact that QS-21 elicits a potent systemic Th1 pro-inflammatory immunity16,17 which caused the AN-1792 side-effects and worsened AD. While most reports about the AN-1792 vaccine assumed that QS-21 was not detrimental to the vaccinated patients, this adjuvant's damaging effects were clearly indicated in an article showing the higher rate of decline in the cognitive functions of AD patients receiving QS-21 alone, ie, 6-7 points in the MMSE, as compared to the decline of 3.5-4 points for the untreated AD population.18
A serious problem with vaccines against Alzheimer’s disease is the lack of information about their composition
Remarkably, several AD vaccines have inflammatory adjuvants to enhance the immunity, apparently unaware of the fact that they induce damaging systemic inflammatory immune responses that favour the production of antibody glycoforms which are effectors of inflammation. Another important oversight is the one in which the vaccine has chemical groups like maleimide, which are highly immunogenic and have detrimental effects on the immune response.4,19 Indeed, because of their high immunogenicity, maleimide groups commonly found in cross-linkers cause epitope-specific immune suppression; still, that chemical group is found in some AD vaccines that have entered clinical studies.
The Affitope ADO2 vaccine
For instance, the Affitope AD02 vaccine from
Paradoxically, a serious problem with vaccines against Alzheimer’s disease is the lack of information about their composition. Indeed, scientific publications on Alzheimer’s disease vaccines often have extensive information about the biological systems and assays used but seldom give detailed information about the structure of the antigen such as the cross-linking agent, or the adjuvant used. This absence of meaningful information about the components of these vaccines hinders any attempt to establish a rational relationship between specific components and induced immune response.
Of significance for vaccine development is that most scientists seem to be unaware of the problems caused by different linkers; an unfavourable situation aggravated by the fact that most commercial linkers usually contain the strongly immunogenic maleimide group. However, these complications caused by a highly abnormal immunogenicity may extend to other groups, like those having cyclic rings with aromatic character.19 Moreover, the maleimide's immune suppression is not limited to peptides, but also extends to carbohydrate antigens;23 a situation that implies that immune suppression may be more acute for weak antigens like self-antigens, eg, amyloid-ß.
That maleimides have an adverse role on immunity raises questions about the validity of some studies. Particularly since they can also affect the assays used to assess immunity. From the regulatory point of view it would be advisable to consider the effect of a vaccine's single components (adjuvant, cross-linker, antigen and carrier) besides those of the final vaccine in the pre-clinical phase. These effects may be difficult to identify by testing the vaccine only.
Regrettably, the AD vaccines' clinical failures have been taken as evidence that this vaccine concept is wrong - a decision that disregards the intrinsic problems linked to virtually all Alzheimer’s disease vaccines. This situation would explain the decision by many organisations to reject research dealing with anti-amyloid vaccines. However, like other conditions where, despite initial failures, researchers persevered until they resolved the problem, AD drug development will benefit if it adopts a similar philosophy.
The development of aducanumab strengthens the use of immunological strategies against aberrant Aß oligomers to prevent and/or delay AD - a notion that AD drug development has recently started to recognise. While the search for novel therapeutic targets is advisable, abandoning promising areas without identifying the sources of the problems is unwise. Actually, the success of cancer immunotherapy took decades of research filled with failures and successes. Thus, AD drug development should follow those lessons in its quest for the development of acutely needed effective products against Alzheimer’s disease.
Dante J Marciani holds a ScD in Biological Sciences and a PhD in Chemistry, and was a scientist at
1. Mcmullin E. Galileo on science and Scripture. In P Machamer (Ed.). The Cambridge Companion to Galileo (Cambridge Companions to Philosophy, pp. 271-347). 1998.
2. Britschgi M et al. Neuroprotective natural antibodies to assemblies of amyloidogenic peptides decrease with normal aging and advancing Alzheimer's disease.
3. Dodel R et al. Naturally occurring autoantibodies against b-amyloid: Investigating their role in transgenic animal and in vitro models of Alzheimer's disease. J Neurosci. 2011;31:5847-5854.
4. Marciani DJ. A retrospective analysis of the Alzheimer's disease vaccine progress - The critical need for new development strategies. J. Neurochem. 2016;137:687-700.
5. Sullivan MG. Aducanumab subanalysis bolsters phase III trials in very early Alzheimer's disease. Neurology Reviews. 2016;24(2):8 (www.neurologyreviews.com).
6. Cummings FL, et al. Alzheimer's disease drug-development pipeline: few candidates, frequent failures. Alzheimer's Research and Therapy. 2014;6:37 https://doi.org/10.1186/alzrt269.
7. Marciani DJ. Rejecting the Alzheimer's disease vaccine development for the wrong reasons. Drug Discov Today. 2017;22:609-614.
8. Hall AM, Roberson, ED. Mouse models of Alzheimer's disease.
9. Head E. A canine model of human aging and Alzheimer's disease. Biochim Biophys Acta. 2013;1832:1384-1389.
10. Selkoe DJ. Soluble oligomers of the amyloid ß-protein impair synaptic plasticity and behavior. Behav Brain Res. 2008;192:106-113.
11. Busche MA, et al. Decreased amyloid-ß and increased neuronal hyperactivity in Alzheimer's models. Nat Neurosci. 2015;18:1725-1727.
12. Kensil C, Beltz GA. Compositions comprising the adjuvant QS-21 and polysorbate or cyclodextrin as excipient. WO1999010008A1, publication date 1999-03-04.
13. Marciani DJ. New Th2 adjuvants for preventive and active immunotherapy of neurodegenerative proteinopathies. Drug Discov Today. 2014;19:912-920.
14. Marciani DJ. Alzheimer's disease: Toward the rational design of an effective vaccine. Rev Neuropsiquiatr. 2015;78:140-152.
15. Hull M, et al. Long-term extensions of randomized vaccination trials of ACC-001 and QS-21 in mild to moderate Alzheimer's disease. Curr Alzheimer Res. 2017;14:696-708.
16. Kensil C, et al. QS-21 and QS-7: Purified saponin adjuvants. Dev Biol Stand. 1998;92:41-47.
17. Marciani DJ. Vaccine adjuvants: role and mechanisms of action in vaccine immunogenicity. Drug Discov Today. 2003;8:934-945.
19. Boeckler C, et al. Immunogenicity of new heterobifunctional cross-linking reagents used in the conjugation of synthetic peptides to liposomes. J Immunol Methods. 1996;191:1-10.
20. Mandler M, et al. Tailoring the antibody response to aggregated Aß using novel Alzheimer-vaccines. PLoS One. 2015;10(1):e0115237.
21. Affitope AD02 | Alzaforum. www.alzforum.org/therapeutics/affitopead02
22. Chackerian B, et al. Virus and virus-like particle-based immunogens for Alzheimer's disease induce antibody responses against amyloid-ß without concomitant T cell responses. Vaccine. 2006;24:6321-6331.
23. Buskas T, et al. The immunogenicity of the tumor-associated antigen Lewis (y) may be suppressed by a bifunctional cross-linker required for coupling to a carrier protein. Chemistry. 2004;10:3517-3524.